3D model-based restoration with positivity constraint using a reduced number of 3D-SIM images

TL;DR

This paper introduces a 3D model-based microscopy restoration method with a positivity constraint, achieving more accurate 3D image reconstruction from fewer raw images compared to traditional approaches.

Contribution

The novel 3D-MBPC method incorporates a positivity constraint via an auxiliary function, improving restoration accuracy with fewer raw images in 3D-SIM data.

Findings

3D-MBPC outperforms previous methods in accuracy.

Effective with reduced raw images (from 15 to 5).

Exploits data redundancy for improved results.

Abstract

We extend our previous three-dimensional (3D) model-based (MB) approach for 3D structured illumination microscopy (SIM) by introducing a positivity constraint (PC) through the reconstruction of an auxiliary function using a conjugate-gradient method. The performance of the new 3D-MBPC method is investigated with noisy simulation and compared to our previous 3D-MB approach and to the 3D Generalized Wiener filter approach used traditionally for 3D processing of 3D-SIM data. Results show more accurate 3D restoration is possible with the 3D-MBPC method over the other two methods. Moreover, information redundancy in 3D-SIM data is exploited and results obtained with the 3D-MBPC method when the number of raw SIM images is reduced from 15 down to 7 and 5 are promising.